Comparison of different classification algorithms for the detection of changes on water bodies; Karakaya dam lake

Optimum management of water and water bodies is crucial in ensuring and maintaining the natural ecosystem cycle. Benefits from wetlands in the world and in our country keep humanity alive. Resources that are of vital importance should be monitored and changes should be observed. Thanks to the science of remote sensing, researchers in many parts of the world can monitor changes in the waters of the earth through satellite imagery and terrestrial supporting studies. The main component of change detection in remote sensing is the classification process. Nowadays, the Classification process has reached different dimensions with the contributions of artificial intelligence and machine learning algorithms. The emergence of different classification algorithms also affected the results obtained from the analyzes. In this study, the change occurred between 1990-2000-2010-2019 in Karakaya Dam Lake, which is included in the borders of Malatya - Elazığ provinces, was observed. In this context, supervised classification processes and change detection analyzes were performed using Landsat satellite data with maximum likelihood, artificial neural network, support vector machine and decision tree algorithms. For detecting the change analysis, the lake boundaries obtained from official sources were used and compared. The data obtained as a result of the study were compared for each algorithm and the amount of change was interpreted.